Cientiic Paper
Brazilian Journal of Applied Technology for Agricultural Science, Guarapuava-PR, v.6, n.3, p.07-16, 2013
Development of technologies and
methods for monitoring the spatial
variability of air temperature in
greenhouse environment
Diego Scacalossi Voltan1
Rogério Zanarde Barbosa1
João Eduardo Machado Perea Martins2
Célia Regina Lopes Zimback3
Received at: 09/04/2013 Accepted for publication at: 09/09/2013
1 Post Graduation student Agronomy - Irrigation and drainage. Universidade Estadual Paulista "Júlio de Mesquita Filho" (UNESP) - Faculdade de Ciências Agronômicas. Botucatu-SP. E-mail: diegosvoltan@gmail.com; rogerio@fca.unesp.br
2 Dr. Prof. Departamento de Computação (FC). Universidade Estadual Paulista "Júlio de Mesquita Filho" (UNESP) - Faculdade de Ciências Agronômicas. Botucatu-SP. E-mail: perea@fc.unesp.br
3 Dra. Prof. Departamento de Recursos Naturais. Universidade Estadual Paulista "Júlio de Mesquita Filho" (UNESP) - Faculdade de Ciências Agronômicas. Botucatu-SP. E-mail: czimback@gmail.com
Abstract
Climatic factors directly inluence growth and productivity of plants inside greenhouses, where temperature can be considered one of the major parameter in this context. Thus, the aim of this research was to develop a low cost device for thermal sensing and data acquisition, and use it in data collection and analysis of spatial variability of temperature inside a greenhouse with tropical climate. The developed equipment for thermal measurements showed a high degree of accuracy and fast responses in measurements, proving its
eficiency. The data analysis interpretations were made from the elaborations of variograms and of tridimensional maps generated by a geostatistical software. The processed data analysis presented that a greenhouse without thermal control has spatial variations of air temperature, both in the sampled horizontals layers as in the three analyzed vertical columns, presenting variations of up to 3.6 ºC in certain times.
Keywords: Spatial dependence; geostatistics; thermometric tool.
Desenvolvimento de tecnologias e métodos para monitoramento da variabilidade
espacial da temperatura em ambientes de cultivo protegido
Resumo
Os fatores climáticos inluenciam diretamente o crescimento das plantas e a produtividade dentro de casas de vegetação, sendo que a temperatura pode ser considerada um dos fatores mais importantes nesse contexto. Assim, o objetivo desse trabalho foi desenvolver um dispositivo de baixo custo para sensoriamento térmico e aquisição de dados, e utilizá-lo no levantamento de dados e análise da variabilidade espacial da temperatura no interior de uma casa de vegetação de clima tropical. O dispositivo físico de medição térmica desenvolvido apresentou um alto grau de exatidão e respostas imediatas nas medições, comprovando sua eiciência. As interpretações dos dados foram feitas a partir da elaboração de variogramas e de mapas tridimensionais gerados por um software geoestatístico. Os dados analisados mostraram que uma casa de vegetação sem controle térmico apresenta variações espaciais da temperatura do ar tanto nas camadas horizontais amostradas, como nas três alturas de colunas verticais analisadas, apresentando variações de até 3,6ºC em determinados horários.
Palavras-chave: Dependência espacial; geoestatística; ferramenta termométrica.
Desarrollo de tecnologías y métodos para monitoreo de la variabilidad espacial de
la temperatura en ambientes de cultivo protegido
Resumen
mediciones, lo que demuestra su eiciencia. Las interpretaciones de los datos se hicieron con el desarrollo de variogramas y mapas tridimensionales generadas por un software geoestadístico. Los datos analizados mostraron que un invernadero sin control térmico presenta variaciones espaciales de la temperatura del aire, tanto en las capas horizontales muestreadas, como en las tres alturas de columnas verticales analizadas, mostrando variaciones de hasta 3,6 ºC en ciertos momentos.
Palabras clave: dependencia espacial; geoestadística; herramienta termométrica.
Introduction
The temperature is one of the most important factors to be observed in the agriculture (TERUEL, 2010; CHEN et al., 2011), since this parameter greatly influences the metabolic functions of the plants and even of animals (SOTO-ZARAZÚA et al., 2011).
In greenhouse environments, AGUIAR et al. (2000), BÖHMER et al. (2008), ROMANINI et al. (2010) and ANDRADE et al. (2011) proved that the influence of the type of cover of greenhouse alter the internal temperature of these environments,
showing that this factor must be taken into account
for substantially affecting the yield.
According to ZHANG et al. (2010), the thermal variation inside protected environments is related to many other factors, among them the solar radiation, convection movement of air masses in the greenhouse edges and by the dynamics of pressure of
air steam which occurs due to a difference of internal
and external temperature of the greenhouse. The advanced control of climate in a greenhouse involves many variables and complicated
non linear processes, which can demand a high
financial investment and a great technological effort to achieved efficient and reliable results (KITTAS and BARTZANAS, 2007; ÖDUK and ALLAHVERDY, 2011; XIU-HUA and LEI, 2011).
Many researchers simplify the study, understanding the interior climate of the greenhouse as uniform (KITTAS and BARTZANAS, 2007; KOLOKOTSA et al., 2010) and thus, usually perform the climatic monitoring through measurements in a unique spot, such as the center of the greenhouse, and the information are then extrapolated for the other spots. This technique is simple for the fact of demanding the physical measurement of only a specific spot.
However, HASAN et al. (2009) and JÁNOS et al. (2010) show that, in the specific case of
the temperature, the creation of a process for measurement in many spots of the greenhouse
allows the making of a map. Inside this concept of
measurements in several places and maps creation,
the use of geostatistical methods appears as an efficient tool for the description and analysis of the spatial distribution of the temperature in greenhouses (SAPOUNAS et al., 2008).
Although there are many commercial models of dispositive for thermal measurement, in this
study we present the development of one of thermal
sensing. Thus, the study has as objective to present an analysis of the spatial dependence of air temperature inside greenhouses using a measurement instrument developed for acquisition of temperature data.
Material and Methods
The development of the low cost dispositive for thermal sensing was based in the temperature sensor LM35, which can be found in the Brazilian market, with cost below to R$ 10.00, it also has a high degree of reliability, with linear output of voltage 10 mV/°C, accuracy of 0.5 °C, power ranging from 4 to
30 V, electric current consumption of 60 µA and auto heating inferior to 0.1 °C. The output voltage can be directly read on a multimeter adjusted in a scale of
millivolts, presenting a relation with the measured
temperature in degrees centigrade (Temp) expressed by Temp = Volt/10.
The methodology of development of the
analysis of spatial thermal variability, where the geostatistics is used as a tool allows quantifying the spatial dependence between the sampled variables and reproduce in graphics how it behaves on the
environment. Thus, in this study the measures of
temperature were sampled in different spots inside
a greenhouse located in the Universidade Estadual Paulista, Faculty of Agronomic Sciences, in Botucatu
SP, with geographical coordinates of 22°51’3’’ of south latitude and 48°25’37” of longitude west and
786 m of altitude.
The representative spots for each air volume
were pre established in the plating lines of the
greenhouse, grid shaped (sampling grid), defined
in 0.8 m between lines and 1.0 m between spots,
8th of 2010 for each of the 114 spots were performed
measurements in three different heights, which were 0.30, 1.20 and 2.00 m. This experiment was done three times a day, at 9, 12 and 16 hours, so that was
established a spatial analysis associated to temporal
variations. The temperature was sequentially
measured in each spot, being that the time of the
total route was 10 min for the measure in each one
of the three heights.
The greenhouse used in this study was of the arch type, 6 m wide, 24 m of length and a ceiling
height of 3 m, positioned in the East-West. The
lateral windows were coated with a polyethylene
mesh of anti-aphid protection. In the superior edges,
longitudinal direction, the greenhouse has two frontal windows with controlled opening to allow ventilation in the warmest times of the day. It also has a polyethylene cover of low density for protection
against rain, from the ridge to the eaves.
The values calculated from the 114 sampled
spots to obtain the descriptive statistic were relevant
to verify the variability of the air temperature data. It
were calculated the average, median, maximum and
minimum value, amplitude, standard deviation and coefficient of variation (CV).
The function of spatial variance
γ
(h),known as variogram allowed to calculate the
spatial dependence by the variance measure of the
differences of the sampled values which are distant
in h m. It is expressed by:
[1]
( )
( )
∑
( )[
( )
(
)
]
= + − = h N i ii Z x h
x Z h N h 1 2 2 1 γ Where:
N(h) is the number of pairs of median values Z(si), Z(si+h), separated by a vector h, being h the
distance between the spots of all the sampled values.
The calculated values for the elaboration of
the semivariograms were: nugget effect or ‘Nugget’
(Co), which represents the value of
γ
(h) when h= 0; Sill (Co + C) is when the value of
γ
(h) stabilizers and its value is approximately equal to the data variance; the Range (a) is the distance h whenγ
(h) achieves the level and the samples become independent. These parameters assist in the analysis of spatial dependence calculated by the relation C/Co + C,denominated structure or spatial proportion which,
according to the classification adapted by ZIMBACK
(2001), if the obtained value of this ratio is ≥ 0.75, it is classified as strong spatial dependence; between
0.25 and 0.75 it is moderate spatial dependence; and
values of ≤ 0.25 as low spatial dependence.
After the analysis of spatial dependence,
the data were interpolated by the kriging method. This method allows estimating values of variables
distributed in the space using the structural properties
of the variogram. The results of the interpolation were
visualized in tridimensional maps representing the spatial distributions of air temperature. All the data
obtained for the analysis were calculated using the
program GS+ (Geostatistical for Environmental Sciences) (Robertson 1998).
Results and Discussion
This section presents the obtained results,
being that, for effects of organization they were divided in two parts which have the discussions related to the low cost dispositive for thermal sensing
and the analysis on the thermal spatial variability.
Low cost dispositive for thermal sensing
Figure 1 exemplifies the electronic circuit and the physical assembling of the dispositive developed
in this study, whose mounting is simple and has the advantage of allowing that the height of the
positions of temperature measurement gets easier and quickly altered. In this study, the dispositive of thermal sensing includes, besides the LM35 sensor, mechanical parts of the probe, the meter for visualization of the measurements and a box of battery conditioning.
The probe was installed in an aluminum tube with 100 cm, being longer than the usual to allow the
easy measurement in different heights and to enable
its insertion in closed places. In one tip was placed a LM35 fixed in epoxy resin and in the other edge was
placed a manual support and a small plastic box for
conditioning of a battery of 9V used to power the sensor. The output of the voltage (Volt) was directly
read by a simple multimeter DT830D model and its display had 3 ½ digits. Thus, in measurement scale of voltage of 2,000 mV, the temperature can
be directly calculated with up to two digits integer and a decimal, which is fairly suitable for numerous
agricultural applications.
Initially was tested a circuit with this physical assembling and without special precautions with the
connection of grounding or shielding. For analysis
of performance of the developed thermometer were made 10 operational tests, being that in each test were
performed 600 temperature measurements, using an
interval of 10 ms between each measurement. Figure 2 shows the graphic of voltage variation of output of the
developed thermometer for the first sampling done.
Table 1 shows the obtained results in the ten samplings, where it can be verified that the values of
the averages and of standard deviation of sampling are very close, proving the stability of the system. In the table, the value of each sample average is
represented in the form of voltage, which can be directly converted in temperature. Table 1 data allows
to calculate the average of the output voltage means
in 269.7 mV, with a standard deviation of 0.264 and with a coefficient of variation of 0.098% and first
confidence interval between 269.436 and 269.964 mV.
With basis in this data, it can be concluded that the developed dispositive of electronic thermal sensing in
this study showed a low cost and excellent accuracy
in the measurements, achieving the initial objective for its development.
Analysis of the spatial thermal variability
Figure 3 shows the temperature variation
and solar radiation in the external part and close to the greenhouse at the day of the measurements,
allowing that the internal data were also analyzed in
Figure 1. Electronic circuit of the thermometer with the LM35 sensor (left) and the actual photograph of the developed thermometer (right).
comparison with external factors, The air temperature presented variations between 17.62 and 28.19 °C during the day, with an average of 24.81 °C, and solar
radiation with values between 852 to 1,029 W/m².
Table 2 shows the statistical variables
calculated through the data registered in the measurement done at the three schedules and three heights of the operation. It can be observed that to the schedules 9h and 16h the coefficients of variation of the temperature averages in the three different
heights was always inferior to 1.8%, however at 12h was critical, being that the temperature within the greenhouse presented peak values with a difference
of 3.6 °C at a height of 0.30 m.
Tables 3, 4 and 5 show, respectively, the
variogram parameters obtained through the geostatic
analysis, adjusted by the software GS+ for the data
of the measures at 9, 12 and 16 h, considering the higher value of regression R². The values assisted in
the elaboration of the variograms and, for it, were
calculated the values of the Nugget effect (C0), of the Sill (C + C0), of Range (a), of regression (R²) and of the structure or spatial proportion [C/ (C0 + C)].
Thus, they were used to calculate the experimental
variograms and then adjust them to the theoretical
models showed in the figures 4, 5 and 6.
At the three schedules, except for the heights 0.30 and 1.20 m at 9h and 1.20 m at 16h, all variograms
were adjusted by the Gaussian model. This model
reveals that the temperatures of the sampled spots
presented continuity among them with certain
regularity. The Gaussian model adjusts to the spots not going through them all and smoothes the curve
when approaching to zero. This model is also used
Table 1. Statistical values of 10 samplings, with 600 samples each, of the output voltage signal of the
thermometer developed with the LM35.
Sampling Sample Average (milivolts) Standard Deviation Coefficient of variation (%)
1 269.039 0.931 0.346
2 269.462 0.944 0.350
3 269.571 0.910 0.337
4 269.654 0.894 0.331
5 269.744 0.902 0.334
6 269.737 0.943 0.349
7 269.744 0.920 0.341
8 269.747 0.937 0.347
9 269.941 0.883 0.327
10 269.961 0.921 0.341
Figure 3. Behavior of the air temperature and of the solar radiation in the external environment in relation to
the greenhouse, in the intervals in which was measured the internal air temperature of the same.
to represent extremely continuous variables and the
value of the Range is correspondent to 95% of the Sill
(Isaaks and Srivastava 1989).
In figure 4 (A) and (B) the exponential model
was better adjusted to the heights 0.30 and 1.20 m
at 9h. This model is evident due to the Sill tends to the infinite, or better, the variance curve is highly
dispersed in relation to the distance between samples.
Despite the elaborated variogram range for the
height 1.20 at 16h being close to the other variograms adjusted to the Gaussian model, it presented a linear
behavior and with a rapid growth in the source,
characteristic of the spherical model, as presented in figure 6 (B).
From the determination of the variogram
parameters was assessed the Range of the samples, parameter which represents the maximum distance
that the spots are related spatially to the same Table 2. Results of the descriptive analysis of statistics of the values collected of air temperature at the schedules 9, 12 and 16h at the heights of 0.30, 1.20 and 2.00 m.
Descriptive
statistics 9 hours 12 hours 16 hours
0.30 (m) 1.20 (m) 2.00 (m) 0.30 (m) 1.20 (m) 2.00 (m) 0.30 (m) 1.20 (m) 2.00 (m)
Average 24.02 25.28 25.55 28.87 28.91 30.62 28.26 31.38 30.06
Median 24.10 25.20 25.55 28.80 29.00 30.50 28.20 31.50 29.90
Maximum 24.70 26.50 26.40 30.60 30.00 32.20 29.50 31.90 31.60
Minimum 23.00 24.70 24.60 27.00 27.80 29.10 27.50 30.60 28.90
Total range 1.70 1.80 1.80 3.60 2.20 3.10 2.00 1.30 2.70
Standard
deviation 0.3865 0.3885 0.4516 0.9619 0.5959 1.0351 0.3615 0.3139 0.4893
CV (%) 1.61 1.54 1.77 3.33 2.06 3.38 1.28 1.00 1.63
Table 3. Variogram parameters set at the time of 9h.
Height (m) Variogram
model
Nugget variance
(Co) Sill (Co+C) Range (m) R
2 C/Co+C
0.30 Exponential 0.0686 0.3242 62.97 0.702 0.788
1.20 Exponential 0.0502 0.3494 62.97 0.930 0.730
2.00 Gaussian 0.0276 0.2302 5.5252 0.919 0.880
Table 4. Variogram parameters set at the time of 12h.
Height (m) Variogram
model
Nugget variance
(Co) Sill (Co+C) Range (m) R
2 C/Co+C
0.30 Gaussian 0.0010 0.9610 2.4942 0.860 0.9990
1.20 Gaussian 0.0010 0.3470 2.4595 0.909 0.9970
2.00 Gaussian 0.0010 1.2760 4.5207 0.896 0.9990
Table 5. Variogram parameters set at the time of 16h.
Height (m) Variogram
model
Nugget variance
(Co) Sill (Co+C) Range (m) R
2 C/Co+C
0.30 Gaussian 0.0230 0.1220 4.9017 0.952 0.811
1.20 Spherical 0.0176 0.1142 5.4700 0.797 0.846
Figure 4. Theoretical models of the variograms adjusted for the 9h schedule, at the heights 0.30, 1.20 and 2.00 m. (A) Exponential, (B) Exponential and (C) Gaussian.
Figure 5. Theoretical models of the variograms adjusted for the 12h schedule, at the heights 0.30, 1.20 and 2.00 m. (A) Gaussian, (B) Gaussian and (C) Gaussian.
Figure 6. Theoretical models of the variograms adjusted for the 16h schedule, at the heights 0.30, 1.20 and 2.00 m. (A) Gaussian, (B) Spherical and (C) Gaussian.
variable and which, according to ANDRADE (2002), marks the distance from which the samples become
independents.
At the 9h schedule, the Range value for the
heights 0.30 and 1.20 was 62.97 m. This distance
reveals that the air temperature is presenting very close values and the variable starts to be independent
starting from this distance. This shows that, possibly,
due to the non direct incidence of solar rays in the
respective heights at this time, in the west part
of the structure, the environment is not suffering direct interference of solar radiation, characterizing a homogeneous environment in relation to the air
temperature. At 2.00 m of height, the solar rays were
reflecting on the greenhouse cover and the top of the plants and the Range distance of air temperature at this height drastically fell to 5.52m, as presented in Table 2.
During the warmest moment of the day,
registered at 12h, the Range achieved the smallest
values on the three sampled categories (0.30, 1.20
and 2.00 m of height) which were 2.49, 2.45 and
4.52 m, respectively. This schedule presented the most significant temperature variations in relation to the 9 and 16h schedules, and possibly its range
was smaller due to the influence of the phenomenon which occurred in isolation inside the greenhouse due
to the high reflection of solar radiation. Still at 12h, in figure 8 is observed that the highest air temperatures occupy the central region of the greenhouse, this variation can be a consequence of the opened lateral of the structure, thus contributing to the dissipation of heat during the moment that the solar radiation is more intense.
At 16h, the Range values for the 0.30, 1.20 and
2.00 m heights were respectively 4.90, 5.47 and 5.14 m. There was a slight increase of the Range in comparison
to the 12h schedule. It can be observed in figure 1 that the air temperature in the greenhouse exterior is decreasing, due to the smaller influence of the solar radiation, and the temperature of the internal air volumes in the greenhouse tends to homogenize, possibly exchanging heat. FURLAN and FOLEGATTI (2002), assessing the air temperature distribution in controlled environment observed that even after the completion of nebulization, at 17h, the heat transmission from the plastic cover to
the internal environment was very low. It can be seen
in table 2 that the total amplitude decreases 0.4, 0.9 and 1.6 ºC in relation to the 12h schedule for the heights of 0.30, 1.20 and 2.00 m.
In this sense, the variograms analysis showed that there was spatial dependence for all times and heights, as shown in the figures 4, 5, 6. It is observed that, according with the index of spatial
dependence adapted by ZIMBACK (2001), except
for the assessment of 9h at the height 1.20 m which
presented moderate spatial dependence, all other values had strong spatial dependence.
The visualization of the spatial dependence
was observed in maps generated by the software GS+ with 3D representation of the spatial distribution
of temperature of the sampled spots at 9, 12 and 16h, obtained through interpolation of the data by
kriging. Figures 7, 8 and 9 show the graphics of spatial
distribution of temperature for each studied schedule at 0.30, 1.20 and 2.00 m of height.
The current study demanded the temperature measurement in several spots of the greenhouse.
Thus, for costs reduction it was used only one datalogger and one probe, with only one sensor. Due to the work conditions, another possibility is to use a single meter with various sampling points in the
same set. Being these possibilities a motivation for
new researches.
Figure 7. 3D Representation of the spatial distribution of air temperature at 9:00h at 0.30 m, 1.20 m and 2.00 m.
Figure 9. 3D Representation of the spatial distribution of air temperature at 16:00h at 0.30 m, 1.20 m and 2.00 m.
References
ANDRADE, A. R. S. Aplicação da Teoria fractal e da geoestatística na estimativa da condutividade hidráulica saturada e do espaçamento entre drenos. Botucatu/SP, Tese (Doutorado em Agronomia) - Faculdade de Ciências Agronômicas, Universidade Estadual Paulista, Botucatu, 2002. 181f.
ANDRADE, J.W. S.; FARIAS JUNIOR, M.; SOUSA, M.A.; ROCHA, A.C. Utilização de diferentes ilmes
plásticos como cobertura de abrigos. Acta Scientiarum. Agronomy. v.33, n.3, p.437-443, 2011.
BÖHMER, C.R.K.; MAUCH, C.R.; ASSIS, S.V.; SCHÖFFEL, E.R.; MENDEZ, M.E.G. Alterações na temperatura do ar proporcionadas por estufa de polietileno, durante um cultivo de feijão-vagem. Revista Brasileira de Agrometeorologia, v.16, n.2, p.143- 148, Ago., 2008.
BOJACÁ, C. R., GIL, R., COOMAN, A. Use of geostatistical and crop growth modelling to assess the
variability of greenhouse tomato yield caused by spatial temperature variations. Computers and Electronics in Agriculture, v.65, n.2, p.219-227, 2009.
BRITO, A.A.A. Casa de vegetação com diferentes coberturas: Desempenho em condição de verão. Tese de Doutorado (Pós-graduação em Engenharia Agrícola), UFV, Viçosa, 2000. 100p.
CHEN, C.; SHEN, T.; WENG, Y. Simple model to study the effect of temperature on the greenhouse with shading nets. African Journal of Biotechnology, v.10, n.25, p.5001-5014, 2011.
DJEVIC, M.; DIMITRIJEVIC, A. Energy consumption for different greenhouse constructions. Energy, v.34, n.9, p.1325-1331, 2009.
EVANGELISTA, A.W.P.; PEREIRA, G.M. Efeito da cobertura plástica de casa-de-vegetação sobre os elementos meteorológicos em lavras, M.G. Ciência e Agrotecnologia, v.25, n.4, p.952-957, 2001.
FURLAN, R.A.; FOLEGATTI, M.V. Distribuição vertical e horizontal de temperaturas do ar em ambientes protegidos.Revista Brasileira de Engenharia Agrícola e Ambiental,, v.6, n.1, p.93-100, 2002.
Conclusions
The air temperature in greenhouse presented strong spatial dependence for all assessed times
and heights, except at 9h and 1.20 of height, which
presented moderate spatial dependence.
The air temperature variations within the
greenhouse had maximum amplitude of 3.6 °C at 12h and minimum amplitude of 1.3 °C at 16h.
The technological resources can be used in a cheap and efficient from for the monitoring and
analysis of temperature variation in greenhouses. The developed electronic thermometer, besides
of the low cost and simplicity of assembling, presented a high accuracy, being considerate satisfactory, which
ensures reliability in the thermal measurements done
with the same.
The detected variations can influence in the uniformity development and in the crops yield, besides of assisting in strategic measures of climatic control, facilitating the management and decision taking.
HASAN, O.; ATILGAN, A.; BUYUKTAS, K.; ALAGOZ, T. The eficiency of fan-pad cooling system in
greenhouse and building up of internal greenhouse temperature map. African Journal of Biotechnology, v.8, n.20, p.5436-5444, 2009.
ISAAKS, E.H.; SRIVASTAVA, R.M. An introduction to applied geostatistics. New York: Oxford University Press, 1989. 560p.
JÁNOS, S.; MARTINOVIĆ, G.; MATIJEVICS, I. WSN implementation in the greenhouse environment using
mobile measuring station. International Journal of Electrical and Computer Engineering Systems, v.1, n.1, p.37-44,, 2010.
KITTAS, C.; BARTZANAS, T. Greenhouse microclimate and dehumidiication effectiveness under different
ventilator conigurations. Journal Building and Environment, v.42, n.10, p.3774-3784, 2007.
KOLOKOTSA, D.; SARIDAKIS, G.; DALAMAGKIDIS, K.; DOLIANITIS, S.; KALIAKATSOS, I. Development of an intelligent indoor environment and energy management system for greenhouses. Journal Energy Conversion and Management, v.51, n.1, p.155-168, Jan., 2010.
MESMOUDI, K.; SOUDANI, A.; BOURNET, P.E. The determination of the inside air temperature of a
greenhouse with tomato crop, under hot and arid climates. Journal of Applied Science in Environmental
Sanitation, v.5, n.2, p.117-129, 2010.
ÖDUK, M. N.; ALLAHVERDY, N. The advantages of fuzzy control over traditional control system in greenhouse automation. In. anais Artiicial Intelligence & Machine Learning Conference. Dubai: UAE, 2011. p.91-97.
SAPOUNAS, AA., NIKITA-MARTZOPOULOU, CH. and SPIRIDIS, A. Prediction the spatial air temperature distribution of an experimental greenhouse using geostatistical methods. In anais: International Symposium on High Technology for Greenhouse System Management: Greensys200, Acta Horticulturae, Naples, v.2, n.801, p.495-500, 2008.
SILVA, M.A.A.E.; GALVANI, E. ; ESCOBEDO, J.F.; CUNHA, AR. Avaliação da temperatura e umidade relativa do ar em estufa com cobertura de polietileno. In. anais Congresso Brasileiro de Meteorologia, 11, Rio de Janeiro, RJ: Sociedade Brasileira de Meteorologia, 2000. 1 CD-ROM.
SONI, P.; SALOKHE, V.M.; TANTAU, H.J. Effect of Screen Mesh Size on Vertical Temperature Distribution in Naturally Ventilated Tropical Greenhouses. Biosystems Engineering, v. 92, n. 4, p. 469–482, 2005.
SOTO-ZARAZÚA, G.M.; GARCIA, E.R.; AYALA, M.T. Temperature effect on ish culture tank facilities inside greenhouse. International Journal of the Physical Sciences, v.6, n.5, p.1039-1044, 2011.
TERUEL, B.J. Controle automatizado de casas de vegetação: variáveis climáticas e fertirrigação. Revista Brasileira de Engenharia Agrícola e Ambiental, v.14, n.3, p.237-245, 2010.
WARRICK, A.W.; NIELSEN, D.R. Spatial variability of soil physical properties in the ield. In: HILLEL,
D. (Ed.). Application of soil physics. New York: Academic Press, 1980. 385p.
XIU-HUA, W.; LEI, Z. Simulation on temperature and humility nonlinear controller of greenhouses. In. anais International Conference on Intelligent Computation Technology and Automation (ICICTA), Shenzhen: 2011, p.500-503.
ZHANG, Z. K.; LIU, S. H.; HUANG, Z. J. Estimation of cucumber evapotranspiration in solar greenhouse in northeast China. Agricultural Sciences in China, v.9, n.4, p.512-518, 2010.
